Neural network based three dimensional ocean modeler

Abstract

A method is described for providing an estimate of the state of a moving tact in a three dimensional ocean. The method comprises the steps of providing a device for estimating the state of the contact, inputting into the device information about a location of an observer during a sequence of time, information from at least one sensor about the position of the contact relative to the observer during the time sequence, and a knowledge vector, transforming the inputted information into a series of three dimensional geographical grids, and analyzing the geographical grids to produce an estimate of the state of the contact with respect to the location of the observer. The device for providing the estimate of the state of the moving contact is a neurally inspired contact estimation device. The device includes a grid stimulation block for transforming the inputted information into the three dimensional geographical grids, a fusion block where grids corresponding to similar time intervals are combined or fused, a correlation block for providing constraints such as constant speed and heading and for producing a path likelihood vector, and an estimate block for providing the estimate of the state of the moving contact. The device further includes a controller for providing knowledge to the aforementioned blocks.

What is claimed is: 
     
       1. A method for providing an estimate of the state of a moving contact comprising the steps of: providing a device for estimating the state of said contact;   inputting into said device information about a location of an observer during a sequence of time, information from at least one sensor about a position of said moving contact relative to said observer during said sequence of time, and a-priori knowledge;   transforming the inputted information into a series of three dimensional geographical grids with each grid having a plurality of cells; and   analyzing said geographical grids to produce an estimate of the state of the contact with respect to the location of the observer.   
     
     
       2. The method of claim 1 wherein said inputting step comprises inputting knowledge including at least one of environmental data, sensor characteristics, contact kinematics and historical data. 
     
     
       3. The method of claim 1 wherein said inputting step comprises: inputting sensor data which includes at least one of location relative to the observer and contact measurements; and   inputting knowledge data which includes at least one of measurement error and confidence in the sensor.   
     
     
       4. The method of claim 1 further comprising: stimulating said cells in said geographical grids in proportion to the likelihood of a contact's presence; and   said analyzing step comprises fusing said geographical grids with said stimulated cells so that grids corresponding to similar time intervals are combined into a series of consolidated grid representations.   
     
     
       5. The method of claim 4 wherein said analyzing step further comprises: applying a constant speed and course constraint for contact motion to said consolidated grid representations; and   producing a path likelihood vector containing a likelihood measure for each possible contact path.   
     
     
       6. The method of claim 5 wherein said analyzing step further comprises transforming said path likelihood vector into said estimate of the contact state. 
     
     
       7. The method of claim 6 further comprising: providing a controller having access to said a-priori knowledge; and   transferring said knowledge from said controller to various functional blocks for performing said information transformation step, said stimulating step, said fusing step, said constant speed and course constraint applying step, said path likelihood vector producing step, and said path likelihood vector transformation step.   
     
     
       8. The method of claim 7 further comprising: monitoring said various functional blocks with said controller; and   adjusting parameters within at least one of said blocks using said controller to improve the contact state estimate.   
     
     
       9. The method of claim 1 wherein said transforming step comprises: forming observation and location pairs for each sensor;   transforming said pairs into three dimensional continuous probability density functions representing the likelihood of the contact's location over the respective time plane;   incorporating at least one of three dimensional sound propagation characteristics, three dimensional sensor motion compensations, and measurement confidence into said probability density functions;   transforming said three dimensional continuous probability density functions into discrete values corresponding to stimulation levels for the cells in said geographical grids;   choosing at least one of a coordinate system and a cell resolution using a controller having access to said a-priori knowledge; and   using said controller to analyze the grids and make adjustments to a set of controlling parameters.   
     
     
       10. A system for providing an estimate of the state of a moving contact which comprises: a device for estimating the state of said contact;   means for inputting into said device information about a location of an observer during a sequence of time, information from at least One sensor about a position of said contact relative to said observer, and a-priori knowledge;   said device having means for transforming said knowledge and said observer and said location information into an estimate of the state of said contact with respect to the location of said observer;   said device having controller means for providing said knowledge to said transforming means; and   said transforming means having a stimulation block for transforming said observer and sensor information into a series of three dimensional geographic grids with each grid having a plurality of cells.   
     
     
       11. The system of claim 10 wherein said stimulation block comprises: first means for transforming pairs of said sensor observation information and said location information into three dimensional continuous probability density functions representing the likelihood of the contact's location over the respective time plane; and   second means for transforming said probability density functions into discrete values corresponding to stimulation levels for the cells in said three dimensional geographical grids.   
     
     
       12. The system of claim 11 wherein said controller means further comprises: means for providing knowledge about at least one of three dimensional sound propagation characteristics, three dimensional sensor motion compensations, and measurement confidence to said first means;   means for advising said second means about at least one of a coordinate system and a cell resolution; and   means for analyzing said grids and making adjustments to parameters controlling said grids.   
     
     
       13. The system of claim 10 wherein said transforming means further comprises an information fusion block for combining ones of said grids corresponding to similar time intervals into a series of consolidated grid representations. 
     
     
       14. The system of claim 13 wherein said information fusion block has a series of artificial neurons for combining like-regions from each geographical grid. 
     
     
       15. The system of claim 14 further comprising said controller affecting cell fusion according to a-priori knowledge and the formation of the consolidated grid representations. 
     
     
       16. The system of claim 13 wherein said transforming means further comprises a correlation block for applying motion constraints to said consolidated grid representations and producing a path likelihood vector containing a likelihood measure for each possible contact path. 
     
     
       17. The system of claim 16 wherein said correlation block has a series of artificial neurons for computing a measure of the likelihood that a particular path was taken by the contact, each said neuron receiving an input from each of said consolidated grid representations. 
     
     
       18. The system of claim 17 further comprising said controller means affecting said neurons according to a-priori knowledge including at least one of maximum speed and depth and depth change characteristics. 
     
     
       19. The system of claim 16 wherein said transforming means further comprises an estimation block for transforming the path likelihood vector to said estimate of said contact state. 
     
     
       20. The system of claim 19 wherein said controller means further comprises: means for advising said estimation block on an averaging method; and   means for providing constraints to be placed on said estimate.   
     
     
       21. The system of claim 16 wherein said controller means further comprises: means for monitoring each said block; and   means for adjusting at least one of parameters and knowledge to improve the estimate of said contact state.